k.) 1990 Oxford University Press

Nucleic Acids Research, Vol. 18, No. 11 3195

Differential expression of jun and fos genes during differentiation of mouse P19 embryonal carcinoma cells Rolf P.de Groot, Jon Schoorlemmer, Siebe T.van Genesen and Wiebe Kruijer* Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands Received March 14, 1990; Revised and Accepted May 11, 1990

ABSTRACT The jun and fos gene families encode DNA binding proteins involved in transcriptional regulation of genes containing a TPA responsive element (TRE). To study their role in gene regulation during early mammalian development, expression and transcription regulatory properties of their gene products were investigated during retinoic acid (RA) induced differentiation of P19 embryonal carcinoma (EC) cells. Our results show, that c-jun is expressed at low but detectable levels in undifferentiated P19 EC cells and at elevated levels in its RA differentiated derivatives, corresponding with increased expression of Jun and TRE binding activity. Jun D is constitutively expressed at constant levels in both undifferentiated and differentiated P19 cells, while jun B and c-fos are not expressed. Addition of TPA to undifferentiated P19 cells does not result in induction of c-jun, jun B and c-fos, while these genes are transiently induced in RA-differentiated P19 cells. In addition, TPA treatment resulted in expression of Fos and Jun protein in RA-differentiated, but not in undifferentiated P19 cells. Addition of TPA to P19 EC cells expressing low levels of TRE binding proteins is neither followed by transcriptional activation of the TRE reporter gene nor by induction of c-jun, previously shown to be autoregulated by its own gene product. By contrast, in P19 cells differentiated by RA that contain elevated levels of TRE binding proteins, TRE dependent transcription is enhanced upon TPA treatment.

INTRODUCTION Embryonal carcinoma cells are the malignant stem cells of teratocarcinoma and resemble the pluripotent cells from the inner cell mass of early mouse preimplantation embryos (1,2). Undifferentiated EC cells are tumorigenic cells, that exhibit all the charasteristics of transformed cells. These include anchorage independent growth and proliferation in the absence of exogenous growth factors in vitro and tumorigenicity in vivo when transplanted to an extra-uterine site. EC cells have the capacity to differentiate in vitro in response to chemical agents, such as *

To whom correspondence should be addressed

retinoic acid (3,4). Their differentiated derivatives are nontumorigenic cells that express growth factor receptors and require exogenous added growth factors for proliferation. The reversible nature of the transformed phenotype of EC cells suggests, that the tumorigenic/growth factor independent growth state of EC cells and the non-tumorigenic/growth factor dependent growth state of their differentiated derivatives results from the onset of different genetic programs that can be modulated by RA. Since RA is mediating its effects through the nuclear retinoic acid receptors (5,6), it is conceivable that characteristic changes in patterns of gene expression are responsible for this effect. In order to obtain more insight in the molecular mechanisms underlying these different growth mechanisms, we have investigated the expression of the jun and fos genes and the involvement of their products in TRE dependent transcription during RA induced differentiation of EC cells. c-Jun is the cellular homolog of vjun, the transforming gene of avian sarcoma virus 17 (ASV 17) and encodes a major component of transcription factor API (7-10). The c-jun gene product binds to the TPA responsive element (TRE) enhancer sequence present in the promoter region of a variety of TPA responsive genes (8,11 13). Jun is capable of forming a heterodimeric protein complex with Fos, which exhibits increased affinity for DNA compared to the homodimeric Jun protein complex (14-19). This is caused by a greater syability of the Fos/Jun heterodimer compared to the Jun/Jun homodimer (68). Site directed mutagenesis has indicated that the Fos-Jun protein complexes associate according to a parallel 'leucine zipper' (20-27). Both Fos and Jun are rapidly induced by growth factors in a variety of somatic cells and are post translationally modified by phosphorylation (28-34). They are members of a larger family that includes the fos related antigen (fra-1), fos B, jun B and jun D genes (35-38). Fos-Jun complexes have been shown to be involved in both positive and negative transcriptional regulation of target genes containing the TRE/API enhancer, including aP2, collagenase, transin, fos and c-jun itself (11,12,39-43). Previous investigations have shown, that F9 EC cells lack functional c-jun/ API activity, while transient transfection of the v/c-jun genes restores transcription from reporter constructs containing a TRE (10,14,15,44,45). Differentiated derivatives of F9 cells express TRE binding proteins, which correlates with the TPA-induced expression of genes containing a TRE (46). In -

3196 Nucleic Acids Research, Vol. 18, No. 11 view of the different growth properties of EC cells versus their differentiated derivatives, we have investigated the differentiation dependent expression of TRE binding protein genes. Our results indicate that the c-jun, jun B and c-fos genes are differentially expressed and induced in both states of differentiation. Implications will be discussed in relation to the expression of target genes containing TRE enhancers involved in growth and differentiation of early embryonic cells.

METHODS Cells and plasmids P19 embryonal carcinoma cells and the differentiated clonal isolated cell lines EPI-7, Mes-1 and END-2 were cultured as described earlier (47,48). cDNA's of c-jun (W. Kruijer et al, unpublished) and jun B (37) were cloned into pSG5 (49) to generate plasmids for expression in eucaryotic cells. The plasmid pMT-fos was derived from pMGH (50) by replacing the XhoIBamHI GH specific fragment for the mouse c-fos NarI-BamHI (51). As probes for hybridization studies, a 1.0 kb PstI mouse c-jun genomic fragment, a 1.5 kb EcoRI cDNA fragment of jun B (37), a 1.7 kb EcoRI cDNA fragment of jun D (38), a 1.5 kb EcoRI cDNA fragment of fra-1 (35) and a 0.8 kb Pstl fragment of v-fos were used. RNA isolation and northern blotting Total cellular RNA was isolated by the guanidine isothiocyanate/caesium chloride method of Chirgwin et al.(52). RNA was denatured for 10 min at 68°C in 50% (v/v) formamide, 2.2 M formaldehyde, 20 mM MOPS pH 7.0, 5 mM sodium acetate, 1 mM EDTA, separated through 0.8% agarose/2.2 M formaldehyde gels, and subsequently transferred to nitrocellulose filters (BA 85, Schleicher & Schuell) in 20x SSC. RNA was immobilized by baking at 80°C for 2 hr under vacuum. Hybridization was performed in 50% formamide, 5 xSSC, 50 mM sodium phosphate pH 6.8, 10 mM EDTA, 0.1% NaDodSO4, 0.1 mg of sonicated salmon sperm DNA per ml, 2 x Denhardt solution (1 xDenhardt solution contains 0.02% bovine serum albumin, 0.02% ficoll, 0.02%

polyvinylpyrrolidone) at 42°C overnight. 32P-labeled probes were generated using a multiprime DNA labeling kit (Amersham). After hybridization and washing, filters were exposed to Kodak XAR-5 film at -70°C using intensifying screens.

Cell labeling and immunoprecipitation Cell proteins were labeled for 45 min following a preincubation for 30 min in methionine-free medium by addition of L-[35S] methionine (500 pCi/ml, NEN) to subconfluent 3.5 cm dishes of P19 EC cells or P19 cells treated for 5 days with RA. The labeled cells were lyzed in 1 ml of RIPA buffer and protein was immunoprecipitated as described previously (28). c-Jun protein was immunoprecipitated using anti-bp55 antiserum (15). Fos/p39 complexes were immunoprecipitated with the anti-peptide M2 antiserum (53). DNA transfection and transient expression assays P19 cells were plated in DF-Bic/7.5% FCS at 4.105W per 50 mm tissue culture dish 24 hr prior to transfection, or in DF-Bic/7.5 % FCS/10-6M retinoic acid at 2.105 cells per 50 mm tissue culture dish 72 hr prior to transfection. Two hours before transfection, the dishes recieved fresh medium. Cells were incubated for 20 hr with calcium phosphate precipitated DNA's (20pg plasmid per

50 mm dish), followed by addition of fresh medium. Twenty four hours later, the cells were harvested followed by measuring CAT activity. 1 pug of pSV2Apap was always included to serve as internal control to correct for possible variations in the transfection efficiency between different cell types. PAP assays were performed as described by Henthorn et al. (54). CAT activity was determined as described by Gorman et al. (55), and was quantitated by liquid scintillation counting of TLC plate 14C spots.

Gel mobility shift assay An oligonucleotide containing the collagenase AP1 binding site (11) was synthesyzed and cloned into pGEM plasmids (Promega). The sequence of the oligonucleotide (19-mer) was 5'GATCTATCTGAGTCAGCAG 3'. The oligonucleotides were exscised from the plasmids by Sau3A and BamHI and isolated from a 10% (w/v) polyacrylamide gel. The cohesive ends of fragments were labeled with 32PdATP and 32PdCTP (5000 Ci/mol) using klenow fragment of DNA polymerase I. Labeled DNA fragments were separated from unincorporated nucleotides by gel filtration using Sephadex G-50 spin columns (Pharmacia) equilibrated in 10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 150 mM sodium chloride. Isolation of nuclear extracts was performed by the mini-scale method described by Lee et al. (56). Protein concentration was determined by the Biorad protein assay according to the manufacturers protocol. The electrophoretic mobility shift assay (EMSA) used is based on the procedures described by Fried and Crothers (57) with slight modifications. Binding was carried out in 20 pl of 10 mM TrisHCl (pH 7.5), 1 mM EDTA, 50 mM sodium chloride, 2 mM magnesium chloride, 10% (v/v) glycerol, 1 mM DTT, 0.1 mg/ml poly-(dI-dC), 10 ng/ml pUC8 plasmid DNA, 0.02% (v/v) nonidet P-40. 10 pg nuclear extract was incubated with 0.1 to 0.5 ng of 32P-labeled DNA fragment at room temperature for 20 min. For some experiments, nuclear extracts were preincubated with a fifty fold excess of unlabeled competitor DNA in binding buffer for 10 min on ice. To stop the reaction, 4 pl of 0.2% bromophenolblue, 0.2% xylenexyanol-F, 25% ficoll was added, and the mixture was immediately loaded onto a 5% polyacrylamide gel containing 50 mM Tris-HCl (pH 8.5), 192 mM glycine, 1 mM EDTA. Electrophoresis was carried out at 2V/cm until the samples had entered the gel, and continued at 7V/cm for 3-4 hr. The gel was fixed in 10% methanol/ 10% (v/v) acetic acid for 20 min, dried, and visualized by autoradiography.

RESULTS Differentiation dependent expression of TRE binding protein genes Previously it has been shown, that nuclear extracts of F9 EC cells are devoid of proteins that bind to the consensus TRE/API enhancer sequence in vitro (10,14,15,44,45). Differentiation of F9 EC cells by RA/dbcAMP results in expression of TRE binding activity and transcriptional activation of promoter-reporter DNA constructs containing the TRE enhancer (46). Since these studies did not address the question, whether the absence of TRE binding activity in undifferentiated F9 EC cells resulted from lack of expression of TRE binding protein genes, we investigated their expression both in F9 as well as in P19 EC cells by Northern

blotting.

Nucleic Acids Research, Vol. 18, No. 11 3197

RA

time(hr) 0

1

4

8 24 48 72 120

c-jun jun B

junD

* **

c-fos actn

*

fra 1 Figure 1. Expression of jun and fos genes during RA induced differentiation of P19 EC cells. P19 EC cells were treated with 10-6M retinoic acid (RA) for the indicated times (hours) followed by RNA extraction and Northern blotting. 15 itg of total RNA was electrophorized in each lane. The filter was sequentially hybridized with c-jun, jun B, jun D, c-fos, fra-l and actin probes (see methods section).

Undifferentiated P19 EC cells express very low levels of the c-jun gene (detectable on overexposed autoradiographs, data not shown), while treatment with RA for 24 to 48 hours, during which the cells become committed to form endoderm derivatives, results in a 10-20 fold increase in the steady state levels of cjun mRNA (Fig. 1). A similar increase in c-jun mRNA was observed during differentiation of F9 cells (data not shown). Since P19 EC cells have a broader differentiation potential compared to F9 cells and express 10 times higher levels of phorbol ester receptors (see next section), this cell line was chosen for further investigation. As shown in Figure 1, continued cultivation of P19 cells in the continuous presence of RA did not result in a further increase in steady state levels of c-jun mRNA. Both undifferentiated as well as differentiated derivatives of P19 cells express the c-jun gene as two transcripts of 2.7 and 3.4 kb. cjun mRNAs of similar size have been found in a variety of ;somatic cells (8,30,31). In contrast to expression of c-jun, the jun B gene is neither expressed in undifferentiated P19 EC cells nor in its differentiated derivatives. Jun D, the third member of the jun gene family (38), is expressed at constant levels during RA induced differentiation of P19 cells. Since the proteins of the Jun family are known to form heterodimers with the proteins from the Fos family, we investigated the expression of c-fos and the fos related antigen fra-1 during differentiation of P19 EC cells. As shown in figure 1, both these genes are not expressed in

undifferentiated nor in differentiated P19 cells. The level of expression of 3-actin, that is not modulated during RA induced differentiation of P19 EC cells, shows that approximately equal amounts of RNA were electrophoresed in each lane. These results show, that c-jun is differentially expressed in undifferentiated and differentiated P19 cells, while jun B, jun D, c-fos and fra-I expression remains unchanged during RA induced differentiation of P19 cells.

Differentiation dependent induction of c-jun, jun B and cfos Activation of protein kinase C by TPA results in transcriptional activation of a variety of genes, whose promoters contain a TRE/API binding site (11,58,59). Among the genes whose transcription is increased in response to TPA is c-jun itself, due to positive autoregulation by a TRE/API binding site present in the promoter region of the c-jun gene (43). Since c-jun is expressed at very low but detectable levels in P19 EC cells, the effect of TPA treatment on induction of the c-jun gene was investigated in P19 EC cells. As shown in Figure 2, addition of TPA to P19 EC cells did not result in induction of c-jun mRNA. In addition, jun B, jun D, c-fos and fra-I expression was not induced by TPA. The inability to induce these genes by TPA is not trivial, as P19 EC cells express phorbol ester receptors (approximately 150.000 receptors/cell) and translocate protein kinase C from the cytoplasm to the plasma membrane in response to TPA (60). In RA differentiated cells, TPA addition results in a relatively small (2-3 fold) increase in c-jun expression after 60 min. (Fig. 2), despite greatly elevated constitutive c-jun mRNA levels and expression of TRE binding proteins (see next section). Although the jun B and c-fos genes are not constitutively expressed in P19 EC cells treated for 5 days with RA (fig 1), their expression is rapidly and transiently induced by TPA in these cells. Maximal levels of induction are observed 30-60 min. after TPA addition, which return to prestimulation levels within 2 hours (c-fos) or 4-8 hours (jun B) (Fig. 2). In contrast, TPA treatment of differentiated P19 cells did not result in an enhancement of jun D or fra-1 expression. These results show, that unlike jun D and fra-l, the c-jun, jun B and c-fos genes are differentially induced by TPA in undifferentiated and differentiated P19 EC cells.

Expression of Jun and Fos proteins To investigate whether the observed differential expression of fos and jun genes is accompanied by differential expression of Fos and Jun proteins, we immunoprecipitated proteins from 35Smethionine pulse-labeled cells using anti-jun (bp55) and anti-fos (M2) antibodies. Undifferentiated and RA-differentiated P19 cells were treated for different periods with TPA and lyzed in RIPA buffer to detect both free Jun and Jun-Fos heterodimers. As shown in figure 3, neither Fos nor Jun proteins were immunoprecipitated from undifferentiated P19 EC cells. TPA treatment did not induce Fos or Jun expression in these cells. In RA-differentiated P19 cells however, Jun is constitutively expressed, while TPA treatment for 2 hours increased the abundance of immunoprecipitable Jun 2-3 fold. Fos is not detected in untreated cells, but accumulates during TPA treatment for 2 hours. These results are in agreement with the increased c-jun mRNA levels following RA-induced differentiation of P19 EC cells (Fig. 1), and the induction of c-jun and c-fos mRNA by TPA in RA-differentiated cells (Fig. 2).

.

3198 Nucleic Acids Research, Vol. 18, No. 11

::JunT1

*

*lo 4Wt

.i..r

IW'

s C"* O**--* **@o

C, tin-l

_b

*B

4b* 4: *t*o* ** *

*

UE. b*

fr.] 1 Figure 2. Induction of jun and fos genes by TPA in undifferentiated and differentiated P19 EC cells. P19 EC cells and P19 cells treated for 5 days with 10-6 RA were stimulated with TPA (100 ng/ml). RNA was isolated at the indicated time points (hours) following TPA addition. 15 Ag of total RNA was electrophorized in each lane and analyzed by Northern blotting.

hi.4J

-

S:,

anti - os

anti-Jun NIS TPA (hr)

.4

p. *92 -

00

4. tm~..gIr

1

F

-- 66

)--- Fos ....45

.-{

P39/Jun

:z

Figure 3. Expression of Fos and Jun proteins. P19 EC cells and P19 cells treated for 5 days with RA (Pl9+RA) were labeled with 35S-methionine followed by lysis in RIPA buffer. Labeling was for 45 min. TPA (100 ng/ml) was added 15 min after (TPA 0.5 hr) or 75 min before (TPA 2 hr) the 35S methionine. Immunoprecipitation using anti-jun bp55, anti-fos M2 and non-immune serum (NIS) was performed as described in the methods section. Autoradiography was for 5 days.

Nucleic Acids Research, Vol. 18, No. 11 3199

Differentiation dependent expression of TRE binding proteins The genes of the fos and jun families encode nuclear DNA binding proteins that bind to the TRE enhancer sequence in vitro (8,17-19). Since c-jun is differentially expressed during RA induced EC cell differentiation, the presence of TRE enhancer binding proteins in nuclear extracts of P19 EC and its differentiated derivatives was investigated using the electrophoretic mobility shift assay (EMSA). As probe a 32p labeled doubled stranded oligonucleotide containing the TRE/API enhancer sequence present in the promoter of the collagenase gene was used (11). As shown in Figure 4A, nuclear extracts from undifferentiated P19 EC cells contain low levels of TRE binding activity, which are increased approximately 10 fold in cells treated for five days with RA. The TRE DNA-protein complex is resolved as a single broad band. The formation of this DNA-protein complex is inhibited by addition of a 50-fold molar excess of unlabeled competitor DNA containing the TRE/APl binding site, but not by heterologous DNA sequences (pUC 18) lacking the recognition sequence (Fig. 4A). The increase in TRE binding observed after RA induced differentiation of P19 EC cells correlates with increased expression of the c-jun gene (Fig. 1). Since jun B is not expressed after RA induced differentiation of P19 EC cells and jun D expression remains constant during the differentiation process, the increase in TRE binding activity is most likely due to increased expression of the c-jun gene. This is in agreement with the increase in p39/Jun after RA differentiation of P19 EC cells as detected by immunoprecipitation (Fig 3). The developmental expression of TRE/API binding proteins is not specific for P19 embryonal carcinoma cells, as a variety

A

of mouse and human EC cell lines exhibit a differentiation dependent expression of TRE/API binding proteins. As shown by EMSA in Figure 4B, TRE/API binding activity is low in nuclear extracts from undifferentiated P19, F9 and Tera 2 clone 13 EC cells, but present at greatly elevated levels in their differentiated derivatives obtained after RA treatment as well as in three clonal isolated differentiated cell lines of P19 EC resembling ectoderm (EPI-7), endoderm (END-2) and mesoderm (MES-1) (47,48). In the case of F9 cells, treatment with dbcAMP alone resulted in a small increase in TRE binding activity compared to the combined effects of RA and dbcAMP, suggesting that the increased TRE/API binding activity is RA and differentiation specific. In some experiments, an additional faster migrating DNA-protein complex was formed. The formation of this complex results from non-specific binding as it is not competed away with excess homologous competitor DNA (data not shown). TPA differentially induces expression from a TRE reporter plasmid in P19 EC and its differentiated derivatives In order to investigate, if the rise in TRE binding activity in P19 cells after RA induced differentiation is accompanied by transcriptional activation of the TRE enhancer, undifferentiated P19 cells and P19 cells treated for 5 days with retinoic acid were transfected with a TRE-tk-CAT reporter plasmid. As shown in Figure 5A, undifferentiated P19 have a low basal level of TRE dependent CAT activity, which is increased 8 fold following RA treatment for 5 days. A tk-CAT construct lacking the TRE enhancer is inactive both in undifferentiated as well as differentiated P19 cells, indicating that transcription from these

B + C-

a.

a.

+: )l I

compettor

I